Among the newest directions of research in the last few years, renewed interest in religion took place in the robotics community. Pioneering this interest, SanTO was the first ever Catholic robot to be made. Its realisation went together with the necessity of revising the theoretical bases that should properly identify what are SanTO and other robots of this kind. Its media coverage and later employment acted as a Pandora’s box of potential ethical problems, which range from blasphemy to incompatibility with science, through dangers of nudging, AI bias, privacy and more. This paper analyses the main issues and how they were addressed, and the future direction of research leads to the development of the new model, CelesTE, aiming to a more pronounced ethical design.